Listen To The Show
Transcript
Welcome to The Prompt by Kuro House, your daily AI update. Today, we’ve got some big moves in AI media, new foundational models from Microsoft, privacy concerns with a popular AI note app, and a deep dive into geopolitical tensions impacting tech. Let’s dive right in.
OpenAI just made its first media acquisition, buying the Technology Business Programming Network, or TBPN. This popular talk show, known as a hub for candid conversations with top tech CEOs like Mark Zuckerberg and Satya Nadella, will keep its editorial independence but now operates under OpenAI’s strategy team. The Wall Street Journal reports TBPN is on track to pull in over 30 million dollars this year, making it a significant media property. OpenAI’s head of AGI deployment, Fidji Simo, says TBPN will help bring AI to the world in a way that clarifies its impact on daily life. CEO Sam Altman even called TBPN his favorite tech show, expecting the acquisition won’t soften their critiques of OpenAI.
Microsoft AI has launched three new foundational models to compete in the multimodal AI space. According to TechCrunch, these include MAI-Transcribe-1 for speech-to-text in 25 languages, MAI-Voice-1 for ultra-fast audio generation with custom voices, and MAI-Image-2 for video generation. MAI-Transcribe-1 is 2.5 times faster than Microsoft’s previous transcription service, and MAI-Voice-1 can generate 60 seconds of audio in just one second. Pricing starts at 36 cents per hour for transcription, 22 dollars per million characters for voice, and 5 to 33 dollars per million tokens for image input and output. Microsoft remains committed to its partnership with OpenAI, even as it builds out its own AI stack.
Heads up if you use Granola, the AI-powered note-taking app for meetings. The Verge reveals that, despite claiming notes are private by default, Granola makes them viewable to anyone with the link unless you change settings. This could expose sensitive meeting details if links are accidentally shared. Granola also uses your notes for internal AI training unless you opt out in the settings, though enterprise customers are opted out by default. So, it’s worth double-checking your privacy preferences if you rely on Granola for confidential information.
WIRED reports a tense situation unfolding as Iran threatens to target major US tech companies in the Middle East. The list includes Apple, Microsoft, Google, Meta, IBM, Tesla, and Palantir among 18 companies warned of potential attacks starting April 1st. This escalation raises serious concerns about employee safety, infrastructure security, and the broader geopolitical impact on tech investments. Despite these threats, many employees at these companies seem focused on domestic issues, while executives remain cautious. The conflict could also affect IPO markets and tech valuations as uncertainty grows.
And finally, a story from WIRED about Polymarket’s ambitious but troubled pop-up bar in Washington DC. This event aimed to blend real-time news, betting on prediction markets, and socializing, but technical glitches delayed the opening and hampered the experience. Despite the hiccups, the event attracted industry insiders including Palantir affiliates and cryptocurrency enthusiasts. Polymarket is rapidly expanding its cultural footprint, even as the legal and regulatory landscape for prediction markets in the US remains uncertain. This episode highlights the growing intersection of finance, technology, and entertainment in AI-driven platforms.
That wraps up today’s edition of The Prompt by Kuro House. From OpenAI’s media foray to Microsoft’s new AI models, privacy pitfalls, geopolitical risks, and the spectacle of prediction markets, the AI landscape is evolving fast. Stay curious and stay informed—we’ll be back tomorrow with more.
